With modem and WiFi on I got 5h 34min, and with both off - 7h 45min. With final battery that should translate to about 9h and 12h respectively.
This progress comes mostly due to runtime power management improvements - we don't suspend the CPU! With WiFi on, the phone was reachable and ssh-able through all this time.
@vancha Yes, the CPU can be woken up by the modem.
Our plan is to work on runtime power management to get it into good shape first, then we may consider using suspend to RAM. There is an ATF branch where it works already, but it currently breaks DRAM frequency scaling, so it's not really worth it as it is right now.
@Alexmitter @agx @dos It would sound like push messages could send some sort of interrupt, to temporarily wake up the cpu?
@Alexmitter @agx @dos right :) to be honest I trust the purism developers 100% in choosing the right way to handle this. I can't wait to get the thing, but it's still great to see how it keeps improving the way it does.
@dos out of curiosity, would a suspended cpu be able to process incoming calls and potentially increase battery life even more? Or is it a deliberate decision to keep it powered on when the screen is off?
Amazing to see how much battery life is improving!